Download to read offline
A keyword analysis between initiating posts and lone posts in MOOC discussion
- 1. A keyword analysisbetween lone posts and initiating posts in MOOC discussions Shi Min Chua @shiminchua CMC and Social Media Corpora 2018, Antwerp Why did nobody reply to my post?
- 2. The CMC setting- MOOCs CoursesOpen OnlineMassive Large number of learners -registered learners: up to 40000 -learners who take part in the discussion: up to 5000 -Free -No pre-requisite -No face-to-face -Asynchronous -various subjects -evolving learning design
- 4.
- 5. Types of Posts InitiatingPost Lone Post Reply Number of Posts Number of Tokens 117863 6162230 32080 2401795 54172 2642195
- 6. Why? â Usenet groups(Burke et al., 2007; Himelboim, Gleave, & Smith, 2009) â Online news commenting spaces (Ziegele, Breiner, & Quiring, 2014) â Distance learning online discussion (Dennen & Wieland, 2007) Overabundance of Lone Posts in CMC â Levelling ground, not constrained by face-to-face conversation (Cavangah, 2007) â Content (Joyce, Anand, Kraut, 2014; Ziegele et al., 2014) â Time of posting (Jeong & Frazier, 2008) â Size & topic (Himelboim, 2008) Stylistic and linguistic features
- 7. Motivation for examining InitiatingPosts vs. Lone Posts â Probability of posting again increased (Joyce & Kraut, 2006) â Join the discussion for interactive purposes rather than cognitive gains (Springer, Engelmann, & Pfaffinger, 2015) â Initiate direct and explicit interaction, âwhich is essential for dialectical, critical discourse and argumentation (Rourke & Kanuka, 2007) âMeaning-making and co-construction (Stahl, 2003) âMOOC learners may feel frustrated if their posts are seldom responded to (Hew & Cheung, 2014)
- 8. Motivation for examining InitiatingPosts vs. Lone Posts â Reflection: Conversation with oneself or learning materials (Laurillard, 2012).
- 9. Keyword Analysis: Cut-offCriteria â Comparing initiating posts to lone posts â the log-likelihood ratio test, p < 0.000000000001 (Flowerdew, 2015) â Keyness indicator (effect size): Bayes Factor > 10 (Wilson, 2013) â Normalized frequency > 5 per 100,000 (McEnery, 2016) â Dispersion Measure, Griesâ Deviation of Proportion (Griesâ DP, Gries, 2008; Lijffijt & Gries, 2012) < 0.30
- 10. Keyword Analysis Comparison: Initiatingposts v.s. Lone posts Number of posts Number of token Abbreviation Initiating posts Lone posts Replies Initiating posts Lone posts Replies Technology 1151 4731 1966 92680 262140 89620 History 4176 8652 10280 330828 567627 459203 Technology 1900 4822 3588 105238 169263 144498 Linguistics 5500 4449 5000 501186 297458 304662 Social Sciences 4841 30792 5542 406832 1652576 265959 Finances 1845 2612 5576 188831 232226 399772 Management 1207 5905 1303 73449 310497 60348 Astronomy 4386 11724 7974 201854 420635 307103 Health 2661 23109 5153 197137 1024804 221502 Sciences 1650 4331 2427 93420 216450 109408 Healthcare 720 8525 1273 70113 572934 63330 Geography 2043 8211 4090 140227 435620 216790 Total 32080 117863 54172 2401795 6162230 2642195
- 11. Analysis of Keywords Step1 â Concordance lines of each keyword â Salient meaning/function (McEnery, 2016) â Biber et al (1998), Wmatrix (Rayson, 2017), Step 2 â Functional grouping of keywords â Interpretations based on groups of keywords â Categorized based on function related to dialogic learning and MOOCs Step 3 â Discourse/conversation analysis of selected keywords (OâKeeffe & Walsh, 2016) Labelling Grouping Case Studies Learners Lone Posts Initiating Posts Keywords 77 70
- 12. Dialogic Expansion â Heteroglosia (Baktin,1986) â Alternative voices â âRepresents the proposition as but one of a range of possible positionsâ â Allows dialogic alternatives â âIn my view the banks have been greedy.â Dialogic Contraction â Monoglossic â Categorical or bare assertion â No reference of othersâ viewpoint â âThe banks have been greedy.â â Martin & White (2004) v.s. â Modals, Hedges, Boosters, Pronouns, Evaluatives (Fairclough, 2003) â Stance and intersubjectivity (Du Bois, 2007)
- 13. Stance Expression Initiating PostsLone Posts Modal expression might, would, could will, need, able Hedging perhaps, seems, sort Quantifier any all, lot, much, every Epistemic expression wonder, wondering aware, understanding, learned âŚthis is perhaps because we tend to⌠âŚThis might mean actually walking⌠âŚI wonder would the microbial diversity also mirror⌠We need to be more exact⌠âŚI need to be ambidextrous⌠...money taken in by a Company is not all down to their own effort, it relies onâŚ
- 14. Stance Expression Initiating PostsLone Posts Booster surely, just, rather, else really, very, definitely, always Mental verbs feel, feeling, think, agree, keen, hope, hoping, looking, forward, enjoy, enjoyed, love Evaluative wrong difficult, easy, excellent, better, interesting, informative, great, important, good, new Negation cannot, ca, n't âŚReally looking forward to learn⌠âŚI enjoyed this course and definitely learned a lot in⌠Excellent range of resources, thanks! I agree with this definition regarding health âŚI really donât see the point ofâŚ
- 15. Pronouns Initiating Posts LonePosts Pronouns he I, my, our, their Indefinite pronouns anybody, anyone everyone Addressing general audience, instead of âyouâ which is used in one-to-one setting (Tagg, 2012) Arbitrary instead of all inclusive (Biber et al, 1998)
- 16. Keyword: anybody ininitiating posts Request for recommendation ââŚSo, does anybody have a good suggestion for a text book on Anaconda, Python and Pandas?â Offering recommendation ââŚSuggest anybody else that may still be receiving 'file not found' for exercise 6 could try the same thingâŚâ Looking for shared experience âIs it just me or does anybody else have issues find <sic> the PwC paper?â Is this a frequent phrase?
- 17. Greetings: Among 4375 instances,975 âHi everyoneâ, 868 âHello everyoneâ Appreciation: 428 âThank(s)⌠everyoneâ Making stances -104 ânot everyoneâŚâ -ââŚwe need everyone to control our daily wasteâŚâ Narrative ââŚespecially when considering the spread of disease and how closely everyone lived in the citiesâŚâ Keyword: everyone in lone posts
- 18. Questions and Request ââQuestion: does anybody knows what kind âŚ?â Keywords in initiating posts: anybody, anyone, question, please, ?, wonder, wondering, why Initiating Posts Lone Posts Discourse Particles please, sorry thanks, thank Punctuation ,âŚ-();?"': !. Meta-language question, article information, course, knowledge
- 19. Connectors Initiating Posts LonePosts Connectors if, or, then, example, e.g. also, and â qualify or elaborate on a proposition by specifying a condition â âŚ.bulldozing the homes and 're building luxury houses in its place. (A fine example can be found in Barnett, the Sweets Way estate). â âŚobesity can be past<sic> on as a learnt behaviour to children through parents. An example of this could be that, a person with depression that comfort eats and does very little else thus causing them to be obese⌠â âŚI need both hands to propel my wheelchair, so (for example) on the Underground, I find it difficult to use the automatic gates because they don't give me time to put my ticket awayâŚ.
- 20. Conclusions âInitiating posts: Questions andhedging to show uncertainty Addressing with indefinite pronouns, so anyone could respond, give agency to others Negation to voice out alternatives âLone posts: Self-mention Appreciation, Emotion, Evaluation Agreement
- 21. Others Initiating Posts LonePosts Comparative terms/relational than, same more Grammatical the, that, there, here, does, did, was, were, 's, on, by am, 'm, have, for, about, with, to Punctuation ,âŚ-();?"': !. Speech act mean, explain, tell, says, say, told, called Verbs in past tense/passive form used, tried, came joined Verbs in present tense/infinite form affects, helps, achieve, work , gain, meet, improve Uncategorized 1, one, two, numbers, missing, following, why, whether like, well, week, main, currently, working, opportunity, education, environment, mind
- 22. Acknowledgement â Dr CarolineTagg, Prof Mike Sharples, Prof Bart Rienties â Leverhulme Trust â Course Providers â shimin.chua@open.ac.uk
Editor's Notes
- #2Â Information packaging Feedback: what is the difference between initiating post and lone post which use the same keywords (explicitly ask for replies) Explicit vs. not intending to ask for reply: manual coding
- #3Â Set the stage for further analysis of the discussion discussion
- #5Â 16712 replies by facilitators 996 in post by facilitators
- #16 Addressing general audience with these unspecified pronouns suggested that learners were trying to engage with other learners in a dialogue in their initiating posts with a personal touch despite not knowing any name, and it could be weird to use âyouâ, which was not a keyword in initiating posts  Yes, and this stands in contrast to personal one-to-one text messaging, for example, in which the most frequent word is âyouâ â Tagg (2012)! 508 instances of âanyoneâ (60%) and 45 instances of âanybodyâ (32%) were used in a sentence ended with a question mark 75 instances of âanybodyâ (53%) were modal, âbeâ verb or auxiliary verbs, suggesting question format. For example, âcan anybodyâ, âdoes anybodyâ, âhas anybodyâ, âdid anybodyâ, âcould anybodyâ, âis anybodyâ and âwould anybody
- #17Â involvement, agency
- #18 Every inclusive quantifier (Biber et al, 1998) Addressing general audience with these unspecified pronouns suggested that learners were trying to engage with other learners in a dialogue in their initiating posts with a personal touch despite not knowing any name, and it could be weird to use âyouâ, which was not a keyword in initiating posts  Yes, and this stands in contrast to personal one-to-one text messaging, for example, in which the most frequent word is âyouâ â Tagg (2012)! 508 instances of âanyoneâ (60%) and 45 instances of âanybodyâ (32%) were used in a sentence ended with a question mark 75 instances of âanybodyâ (53%) were modal, âbeâ verb or auxiliary verbs, suggesting question format. For example, âcan anybodyâ, âdoes anybodyâ, âhas anybodyâ, âdid anybodyâ, âcould anybodyâ, âis anybodyâ and âwould anybody
- #20 Addressing general audience with these unspecified pronouns suggested that learners were trying to engage with other learners in a dialogue in their initiating posts with a personal touch despite not knowing any name, and it could be weird to use âyouâ, which was not a keyword in initiating posts  Yes, and this stands in contrast to personal one-to-one text messaging, for example, in which the most frequent word is âyouâ â Tagg (2012)! 508 instances of âanyoneâ (60%) and 45 instances of âanybodyâ (32%) were used in a sentence ended with a question mark 75 instances of âanybodyâ (53%) were modal, âbeâ verb or auxiliary verbs, suggesting question format. For example, âcan anybodyâ, âdoes anybodyâ, âhas anybodyâ, âdid anybodyâ, âcould anybodyâ, âis anybodyâ and âwould anybody